Introduction
Lung neuroendocrine neoplasms (LNENs) account for one fifth of all pulmonary malignancies and comprise four histological subtypes with different clinical and biological characteristics [
1‐
5]. Pulmonary carcinoids represent 10% of LNENs and comprise two major subtypes: typical and atypical carcinoids. Typical carcinoids are highly differentiated tumors that are predominantly curable by surgical resection. Accordingly, the five-year overall survival (OS) rate of these patients exceeds 80% and the recurrence rate is low [
6]. Meanwhile, atypical carcinoids (ACs) are moderately differentiated, intermediate-grade tumors with greater metastatic potential compared to typical carcinoids, and with a five-year OS rate of 50%. As a result of their aggressive nature, ACs often necessitate adjuvant chemotherapy after surgical resection [
7‐
9]. Large cell neuroendocrine lung cancer (LCNEC) is a poorly differentiated, high-grade tumor with a complex biology that shares similarities with both small cell lung cancer (SCLC) and non-SCLC (NSCLC). Therapeutic approaches often overlap with management protocols for SCLC, the most lethal lung carcinoma [
10]. SCLC is a heterogeneous malignancy characterized by genomic instability, early metastasis, and a rapid proliferation rate. As this tumor type is often diagnosed at an advanced stage, curative-intent surgery is rarely performed, and treatment usually consists of chemo-immunotherapy with or without radiation [
11,
12].
Targeted therapy and immunotherapy have revolutionized the management protocols of NSCLC patients, yet the therapeutic advancements in LNENs are poor [
13‐
17]. This is primarily due to the lack of targetable driver mutations and to the conflicting results of immunotherapy-related trials [
12]. The tumor immune microenvironment (TIM) has been shown to play a key role in the efficacy of immune checkpoint inhibitors (ICIs) [
18,
19]. In this context, assessment of TIM, characteristic immune checkpoints, and specific immune patterns of LNENs is an essential step in understanding and improving the efficacy of currently used and forthcoming immunotherapeutic approaches.
In addition to the well-studied immune checkpoint molecules such as programmed cell death protein 1 (PD-1), programmed cell death-ligand 1 (PD-L1), and cytotoxic T lymphocyte-associated protein 4 (CTLA-4), other molecules with relevant functions in antitumor immunity are worth investigating [
20‐
23]. One of these molecules of potential clinical importance is the V-domain Ig suppressor of T cell activation (VISTA), a transmembrane protein that inhibits the effector function of T cells. VISTA is usually highly expressed in tumor-infiltrating lymphocytes, leading to a decreased antitumoral immune response [
24]. High VISTA expression has been described in various malignancies such as melanoma, NSCLC, and pleural mesothelioma [
25‐
27]. OX40L (CD252) is the ligand of the OX40 (CD134) receptor and is usually expressed by antigen-presenting cells (APCs) such as dendritic cells or macrophages [
28]. Importantly, some studies have highlighted that agonists of OX40 and OX40L can enhance antitumoral immunity [
29]. Glucocorticoid-induced TNF receptor (GITR) is also a transmembrane protein and plays a pivotal role in the regulation of effector T cells. Importantly, its activation can regulate antitumoral immune response [
30,
31]. T cell immunoglobulin and mucin domain 3 (TIM3) is an immunoregulatory protein of T lymphocytes, myeloid cells, and several tumor cells (TCs) (e.g., melanoma, breast, and kidney cancer). Since TIM3 promotes the development of several tumors by suppressing antitumoral immunity, blockage of the TIM3 pathway might be a promising therapeutic approach [
32,
33]. Although the druggability of these novel immunotherapy targets has already been validated in preclinical settings, their therapeutic relevance has not yet been assessed in LNEN patients [
28,
34‐
37]. Nevertheless, elucidating their expression pattern in human LNENs should be among the first steps in planning specific clinical trials evaluating the efficacy of particular immunotherapeutics directed against VISTA, OX40L, GITR, and TIM3.
In order to provide insights into the applicability of immunotherapy in highly malignant LNENs, the current study aimed to investigate the expression levels and distribution patterns of immunologic markers of potential therapeutic relevance in SCLC, AC, and LCNEC patients. Notably, to provide a comprehensive overview of immunologic marker expression within the whole tumor, the study was conducted using surgically resected specimens.
Discussion
The last twenty years of lung cancer research have clarified that the immune infiltration of the tumorous and peritumoral regions significantly influences the fate of malignant lesions [
41,
42]. Hence, gaining a better understanding of the TIM and the interactions between tumors and ICs is essential for improving the efficacy of targeted therapies and immunotherapy. As for the latter, the recent advances in immunotherapy in other tumor entities have not yet provided an equal breakthrough in LNENs [
43]. A comprehensive understanding of the TIM is, therefore, crucial for efficacious treatment protocols in the future [
44,
45]. Here, we investigated the expression pattern and clinicopathological relevance of four novel immunotherapy targets in surgically resected LNENs.
Currently, only a few predictive and prognostic markers have been identified in neuroendocrine neoplasms. Orthopedia homeobox protein (OTP) is a promising marker for pulmonary carcinoids. Indeed, OTP and the adhesion molecule CD44 have been described as potential prognostic markers; nevertheless, their exact impact on OS and immunotherapeutic efficacy is still controversial [
46‐
48]. In contrast, mutation in the MEN1 (multiple endocrine neoplasia 1) gene has clearly been shown to be a poor prognostic factor in pulmonary carcinoids [
49]. LCNEC has recently been classified into two major groups based on genomic and transcriptomic levels. LCNEC I is characterized by mutations in retinoblastoma protein 1 (
Rb1) and tumor protein 53 (
TP53), similar to SCLC. In contrast, LCNEC II exhibits alterations in NSCLC-type serine/threonine kinase 11 (
STK11), Kelch-like ECH-associated protein 1 (
KEAP1), and Kirsten rat sarcoma virus (
KRAS); these are commonly seen in lung adenocarcinomas too. These molecular findings support the theory that LCNEC is a “mixed basket” of tumors with different origins [
50‐
52]. SCLC is a particularly aggressive entity and was long considered a homogenous tumor. Recent data, however, demonstrated that SCLC tumors could be sub-segmented into distinct molecular subtypes based on the expression pattern of specific transcription factors (achaete-scute homologue 1 (ASCL1), neurogenic differentiation factor 1 (NEUROD1), POU Class 2 homeobox 3 (POU2F3)) and inflammatory characteristics. These subtypes are characterized by varying morphology, therapeutic responsiveness, and prognosis [
12,
53‐
55].
VISTA is a membrane protein, expressed usually by myeloid cells, granulocytes, and T cells, and it functions as a negative checkpoint ligand for antigen-presenting cells and T cells [
56]. Previous studies concerning other malignancies (e.g., lung, kidney, colorectal, endometrial and ovarian cancers) have described VISTA to be expressed by lymphocytes in the tumor microenvironment as well as by the TCs [
57,
58]. Its prognostic significance is rather controversial, since high VISTA expression is associated with improved OS in epithelioid mesothelioma, but with worse survival outcomes in colorectal tumors [
59‐
61]. In the present study, VISTA expression did not have a significant impact on OS. However, ICs in LCNECs and SCLCs expressed VISTA to a greater extent than in AC tumors. We postulate that VISTA might be highly expressed in malignant lesions as it inhibits T lymphocyte function, and therefore, reduces the antitumor response. Thus, VISTA might be an effective treatment target in these highly malignant lesions. In fact, experiments in murine models have confirmed that VISTA inhibition increases the number of T lymphocytes and boosts their function in the tumor environment [
56]. A phase 1 clinical trial is currently investigating the efficacy of an anti-VISTA monoclonal antibody (JNJ-61610588; NCT02671955) in various solid tumors. Meanwhile, another ongoing multicenter study examines the long-term effects of CA-170, a PD-L1/PD-L2 and VISTA inhibition in solid tumors and lymphomas [
56,
62‐
65].
The OX40 ligand “OX40L” is an immune checkpoint modulator primarily expressed on activated APCs, dendritic cells, B cells, and macrophages. [
66] The attachment of OX40 and OX40L enhances the survival of CD4
+ and CD8
+ cells which increases tumor-specific responses of effector T cells in tumors and neutralizes the suppressive effects of
Treg cells. [
67] A recent study on NSCLC found that elevated OX40L expression is associated with higher CD4
+ infiltration and increased OS. [
68] Studies in SCLC, melanoma, and pancreatic ductal adenocarcinoma have yielded similar results. [
69,
70] Our study did not show any significant survival benefit concerning OX40L expression, which might partly be explained by the low number of included cases. However, we found that AC TCs typically expressed OX40L at lower levels than other LNEN TCs. Interestingly, in our previous study [
24] also conducted on LNENs, we did not find significant differences in OX40 expression concerning histological subtypes [
24]. This might be suggestive of independent OX40 and OX40L expression patterns. Importantly, recent in vivo studies suggest that agonistic and antagonistic therapy toward OX40-OX40L interaction might be a potential therapeutic option [
28,
66,
67].
Another attractive target for immunotherapy is GITR, a costimulatory cell surface receptor [
71]. It is expressed mainly on T cells and NK cells and plays a major role in the activation of effector T cells, thereby making it an appealing target for antitumoral therapy. Indeed, triggering GITR enhances the immune system’s antitumoral response in murine models by stimulating T lymphocyte activity and inhibiting Treg cells. The first clinical trial in which the GITR agonist TRX518 was applied in solid tumors commenced in 2018. Unfortunately, the survival benefits were modest, and the primary endpoints were not met despite combining TRX518 with PD-1 and PD-L1 inhibitors [
30,
31,
72,
73]. Nevertheless, several additional studies investigating GITR targeting are still ongoing [
74,
75]. Increased GITR expression has been shown to be a positive prognostic factor in endometrial carcinoma and head and neck tumors, whereas the opposite has been confirmed in renal carcinoma [
72,
73,
75]. In our cohort, GITR was expressed to a greater degree by AC TCs than LCNEC and SCLC TCs, and ICs in ACs expressed significantly less GITR compared to LCNEC and SCLC samples. Of note, the effects of GITR expression on tumor-infiltrating ICs also vary depending on the elapsed time. Specifically, while GITR activation initially inhibits the Treg cells and thus contributes to increased immune infiltration, activating GITR excerpts an opposite effect as time passes and inhibits the antitumor immune response [
31,
34,
35,
76,
77]. Indeed, we also found in our previous study that AC tumors (where GITR expression is expected to be the highest) had considerably lower levels of tumor-infiltrating CD8
+ and CD3
+ lymphocytes than LCNEC or SCLC tumors with high GITR expression [
24]. In terms of survival, low GITR expression in the tumor environment showed a borderline significant trend for improved survival; this trend remained similar in the Cox-regression model.
TIM3, a negative T cell regulator expressed mainly on NK cells and macrophages, was the last protein examined within the framework of the current study. Free TIM3 promotes T cell activation, but once activated, it inhibits the antitumor immune response, induces immunosuppression, and contributes to poor prognosis [
32,
33,
36]. Therefore, TIM3 might emerge in the future as a promising target for inhibition [
75]. Indeed, blockage of both TIM3 and PD-1 has been demonstrated to result tumor regression in preclinical models, and several clinical trials are currently exploring TIM3 inhibition in solid tumors [
32,
33,
78,
79]. Interestingly, both TCs and ICs expressed TIM3 to a greater degree in ACs than in LCNEC and SCLC tumors. As for its clinical relevance, high TIM3 expression by TCs and ICs tended to be associated with improved survival. Nevertheless, patients with AC tumors generally had better prognosis than those with other LNENs, and TIM3 expression could not be validated as an independent prognosticator in our multivariate model. Still, its high expression in AC tumors makes TIM3 a potential subtype-specific immunotherapeutic target for AC patients.
Despite the evident differences in their expression pattern, cluster analysis of VISTA, OX40L, GITR, and TIM3 expression did not differentiate LNEN subtypes. Of note, however, when supplementing the current results with our findings from an overlapping cohort [
24], it became apparent that LNEN tumors have widely different immune phenotypes. One particularity of this finding is that AC tumors are the least immunogenic entities among all investigated LNENs, albeit they have the highest TIM3 and GITR TC expression. These immune profiles can aid in diagnosing each histologic subset and predicting potential therapeutic responses to immune checkpoint blockade. Nevertheless, as highlighted before, VISTA, OX40L, GITR, and TIM3 expression regulate the tumor immune microenvironment through a series of complex and time-dependent processes [
27,
28,
30‐
33,
35‐
37,
62,
76,
79]. These aspects should be considered when assessing their impact on intratumoral immune cell distribution and antitumor response.
Our study has some limitations that must be addressed in future prospective settings. A small fraction (4,9%) of surgically resected tissue samples were older than 15 years. While most antigens in FFPE blocks are well preserved over time [
80,
81], decreasing nuclear immunosignal intensity might occur in some of these older blocks. Of note, however, during our quality check, we obtained positive staining with routine diagnostic antibodies (CD56 [
38] and Ki-67[
39]) even in the three oldest blocks, and we found no statically significant differences between the older and newer FFPE blocks concerning immunotherapy target expression. Another limitation was that clinical and follow-up data were not available in all cases due to the study’s retrospective nature. This hindered subsequent analyses and precluded in-depth measurements (i.e., cancer-specific survival). Albeit this lack of data did not influence our findings concerning the expression pattern of investigated markers, all results deriving from clinical data analysis warrant further validation. Furthermore, although a relatively large number of rare tumors was collected, the cohort size was still small to reach statistical significance in some instances. Lastly, the current study is rather descriptive and hypothesis-generating than evidence-based since the direct effects of immunotherapeutics could not be assessed. Accordingly, the key immunologic drivers of marker expression could not be assessed within the framework of the current study and need to be verified in future experimental settings. Of note, these limitations were partly counterbalanced because all analyses were conducted on surgically resected specimens, thus avoiding the distorting effects of intra-tumoral heterogeneity and obtaining a complete overview of the tumors’ immunologic landscape.
By investigating the expression pattern of potential immunotherapy targets in intermediate- and high-grade LNENs, the current multicenter study aimed to aid the future implementation of novel immunotherapeutic approaches. We report that high TC TIM3 expression is characteristic of AC tumors, whereas elevated TC GITR levels could be found in both ACs and SCLCs. OX40L expression by TCs is the highest in SCLCs and the lowest in ACs. IC infiltration is the least pronounced in AC lesions, and IC VISTA and GITR expressions are also considerably lower in these intermediate-grade malignancies. Altogether, these results might aid in designing clinical trials evaluating the efficacy of particular immunotherapeutics directed against VISTA, OX40L, GITR, and TIM3 in these hard-to-treat malignancies.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.